I set an exam question recently on the contribution of market research to a successful real estate development. Let’s pass over the issue of what exactly constitutes a successful real estate development – most of the students did! In updating the lecture on the topic, I had a good rummage around Google and Google Scholar to see what I could find on market research for real estate development. It was pretty thin. The real estate development textbooks tend to have rather descriptive sections describing measures of supply and demand in real estate markets. Revealingly, probably the best analytical piece (page 23 onwards but the rest is good too) that I found was written by James Graaskamp in 1981 who invariably had interesting takes on things real estate.
Reading the various answers to the examination question raised lots of questions for me. What exactly does ‘research’ involve in the context of real estate development? How much does the quantity of research vary with the scale and sector of development? How much do the techniques and data applied vary with the scale and sector? Is it fundamentally any different from market research in any other business sector or to what extent do all business sectors have their own specific requirements, technical and data issues? How important is ‘formal’ market research to developers and how does this vary between developers? Do different elements of market research have different importance to developers? Assuming that they do market research, how do developers procure the different types of research that they perceive that they require and who provides it? What do they think of the market research that they get and what do they do with it? To what extent is it prone to problems of bias in the same way that many professional services are? To what extent do developments go wrong because of poor market research? How are advances in data analytics changing the nature of market research in real estate development?
So that it’s clear what exactly is meant by it, it’s worth defining market research? It’s the kind of thing that you assume that you know. One definition that I have seen seems to encapsulate it in three activities:
- identification of a specific market and measurement of its size and other characteristics;
- identification of a need or want and the characteristic of the good or service that will satisfy it; and
- identification of the preferences, motivations, and buying behaviour of the targeted customer.
However, this definition from the Business Dictionary breaks down the above into market, product and consumer research respectively under the umbrella term of ‘marketing research’. It’s probably best not to get too hung up on terminology. The three areas seem to capture what market research is widely understood to be. What’s market research likely to be used for?
Well, to coin a (perhaps glib) phrase – to develop is to decide. Market research can provide an evidence base for making decisions on a development project – for site search and then for forming, designing, evaluating, modelling, timing, justifying, funding, pricing and marketing the development project. It is often referred to – not least by Graaskamp – as part of the development risk management process. It seems essentially to be a process of gathering, generating and interpreting evidence on supply and demand patterns with the objective of using it to make better development decisions. Sound reasonable?
In a real estate development context, it’s worth thinking about what the perfect market research exercise would produce in order to appreciate how challenging the task is. The perfect piece of market research would (?)
- Accurately predict the amount of space that will be demanded at completion of the project
- Accurately predict the amount of space that will be supplied at completion of the project
- Accurately predict the quality and type of space that will be demanded and supplied at completion of the project
- Accurately predict the price of space at completion of the project.
‘Right’ is, of course, subjective here. Tenants tend to like oversupplied markets. There’s all sorts of issues here about equilibrium pricing, perfect markets, normal profits etc. but let’s leave that for proper economists. Graaskamp placed it in an ethics framework proposing the concept of “most fitting use” – development that best fitted the needs of the user, the community and the developer.
The developer must then decide whether to develop. Of course, it is standard for a business to want to know whether there will be a market for their product when it comes to market. If it is a new category of product, there is no established market. At least, as a mature product in a mature market, real estate developers know that businesses and consumers buy real estate.
There are lots of standard market signals about demand conditions and whether a real estate market is currently oversupplied or undersupplied. For offices, retail and industrial, the core market-level indicators tend to be yield trends, rental growth, vacancy rates and absorption rates. In residential markets, the key signals tend to be price changes and time on market. Projections for population and economic growth can help to provide indicators of future demand patterns.
Of course, for developers the key risk is that current high demand is not there when the development is completed or that expected future demand is not there at completion or that competitors have satisfied the demand first. Prices may be lower, selling periods may be longer etc. So, in order to address what the developer ideally wants to know, the market researcher needs to be a market forecaster. Most major investors and advisors produce forecasts of future demand and supply conditions in major markets in order to forecast future rents based on expected future supply and demand trends.
We can’t then avoid the issue of forecast uncertainty. Given that many forecasting processes are based on econometric models, it’s worth re-stating Hendry and Clements (2003, 303) – “all econometric models are mis-specified, and all economies have been subject to unanticipated shifts”. Basically, econometric models can’t explain everything that happens in a market and there will invariably be shocks – sometimes major shocks. In terms of deciding whether, when and how much space to develop, the market research process can be data-driven, forecast-reliant, highly technical and quantitative – and still wrong. This is unavoidable. Any estimates will be prone to error due to data uncertainty, model uncertainty, shocks etc. The issue is how much error is involved rather than whether any error is involved. Not surprisingly, we can’t see the future. That’s not to say that market research can’t provide a useful approximation of what future market conditions will be like. We don’t know much about the extent to which it does.
Like any other forecast, market research in real estate development will be prone to obsolescence. It essentially produces snapshot estimates at a certain point in time. After a piece of research has been completed, new information will emerge and new expectations will form – and new estimates could be generated. So, market research may need to be regularly updated. This obviously is a further cost. Also, it can be more difficult to update a development project that has been designed, is approved and that may even be under construction than it is to update the research. Albeit certain types of development project have more optionality than others in the extent to which they can be changed, stopped or slowed down after implementation has started.
Typically this part of the market research process comes down to estimating the likely gap between space supplied and space demand. Models for this type of exercise tend to be most mature in the retail sector. This is mainly because policy makers have been unwilling to let the market decide how much space should be supplied in this sector and have generally tried to co-ordinate supply. This is due to their wariness (on the community’s behalf) of damaging the health of town centres through oversupply. As a result, for decades they have tended to demand robust justification that there is or will be demand for proposed additional retail space.
Broadly a Surplus/Deficit Analysis is carried out which identifies the total catchment area retail expenditures are identified based on the area’s effective buying income. Capture factors are applied to the total area retail expenditures to determine the amount of total supportable retail space in the catchment area. The amount of supportable space is subtracted from existing space to determine whether a surplus or deficit exists. Similar procedures can be applied in office, industrial and residential markets. However, ‘feeding’ such calculative models is extremely data and assumption intensive and uncertainty intensive. In the retail sector, it’s mainly carried out by retail market modelling specialists located in the research departments of the retailers themselves and real estate and planning consultants.
Typically the output of such models tells us whether there is a market gap and estimates its size. However, it doesn’t tell us what the best locations are likely to be, what the use mix should be and what type of scheme should be built. The developer must decide what use and mix of space to develop and where to develop it at the local level. The focus of the type of research outlined above is mainly on the market rather than a location. The nature of the specific project is not its focus. Use and mix decisions involve zooming into the project rather than the market. Graaskamp distinguished between market data and merchandise data here.
If the estimates suggests that a market gap is exist, deciding on the use can be fairly straightforward and tends to be determined by the use that generates the highest profits. Ignoring planning regulation for the moment…whilst there may be good reasons to include other uses, it is usually optimal to build as much of the most profitable use as it is believed that the market can absorb. However, for certain parts of a site, the most profitable use may be an ancillary use. For instance, in multi-storey residential schemes in central London, the ground floor, street-facing space is often most profitable as retail. It may also be necessary to include some ancillary uses to enhance or support the values of the dominant use. For example, a high quality gym may be required in an apartment development, large offices can have a range of ‘support’ facilities such as restaurants, gyms, creches etc. I suspect that the research underpinning these types of project-level decision is less formal. In conjunction with input from close-to-market architects, designers and agents, financial appraisal models inform decisions about the mix of uses.
The developer must decide on the design of the real estate development. Design has become a fairly ubiquitous and important variable in the production and consumption of many products. For real estate developments, it will be implemented by the architect and contractor. Their focus is on building design in terms of form, image, function and symbolism. Owners and users can obtain utility from positive aesthetic perceptions. The functional aspects of design inside the building (internal appearance, internal finishes, services, facilities and layout) will also affect its demand and be transmitted to the rental and capital prices of the asset. Any experienced agent can point to examples of schemes that took longer to let or had lower rents because of the wrong size of units, insufficient car parking, inadequate provision of lifts etc.
It is at the project scale that qualitative research is likely to become more important. Such research may be fairly light touch and informal guided by expert opinion e.g. from agents, about the preferences and requirements of tenants or buyers. However, it can also be more systematic involving more formal, (often qualitative) data-supported research on user preferences and requirements.
Both market and merchandise focused research may also been augmented by research on trends in the user market. Structural shifts within retail, office and industrial sectors driven by the interaction of technological innovation and adoption, demographic change, climate change, globalisation etc. will be transmitted to the location, level and nature of demand for space. For example, the shift towards smaller, central supermarkets, the growth of the discount sector, omni-channel retailing, sustainability, changes on working patterns, warehouse automation and robotics, automation of professional services, big data, ageing workforce, clustering of creative classes etc. – all are changing the patterns of production and consumption and, in turn, affecting the demand and supply in the space markets.
Often the developer must also convince others that the development project should be permitted, authorised, funded and/or bought. Market research may be an important evidence base here and moral hazard issues can come to the fore in this context. To what extent can market research be trusted given the potential incentives for it to be biased in some circumstances? I’d be amazed if it was always independent given the economic dependence of the many producers of market research upon clients who have an interest in having their project to be justified. The need for judgement on model, data, assumptions etc. provides scope for defensible disagreement between researchers.
Stakeholders can be both internal and external. Key external stakeholders are the community whose interests are supposed to be represented by professional planners and political representatives. As noted above, sometimes they will require an evidence base in order to decide whether to permit a specific project. The market research may be required to demonstrate that there is demand for the proposed use. Alternatively, market research may be required to demonstrate that demand for the current or an alternative use is already being adequately supplied.
Market research is also used for land use planning policy formation. In England and Wales, the use of SHMA (Strategic Housing Market Assessments – I would have liked it to abbreviate to SHAM) and SHLAA (Strategic Housing Land Availability Assessments) are good examples here. There’s some useful literature on their methods and applications. However, this is market research for land use policy formation on regulating real estate development. I suppose that’s also a kind of development decision.
Capital providers (lenders and investors) may require an evidence base on the commercial viability of projects as part of their due diligence processes. Going back to the beginning of the discussion, some development organisations will require a robust evidence base as part of their internal governance and risk management systems. Finally, it may be necessary to provide an evidence base to demonstrate to potential tenants e.g. to secure pre-lets to retailers, that their businesses will be viable in the proposed project.
In this (embryonic?) era of ‘surveillance capitalism’, it would be surprising if big data was not beginning to start to play a larger role in market research in real estate markets generally and in real estate development specifically. In terms of market research and marketing, much of the current commentary tends to be on the residential sector and the use of public and private databases to identify consumers’ preferences and behaviours. It’s difficult as yet see how data intensive, predictive analytics can add value in the thinner, more lumpy, more private and dispersed commercial real estate development sector. However, this may well be a lack of imagination or knowledge on my part.
Given that Graaskamp identified market research as “the key to development”, its applications and effectiveness in practice in real estate development projects are fairly undocumented. Academic researchers who specialise in real estate development tend to be thin on the ground and academic researchers that have worked on market research in real estate development are elusive. I’m not aware of any empirical academic work in this area. There seems to be a gap in the market.